Both ketones showed strong antifungal activity, which was lost with the addition of exogenous surfactin. Whole-genome analyses indicate that mutations in ComQPXA quorum-sensing system, constituted the genetic bases of post-ST conversion, which rewired B.
B. subtilis stably changes phenotype against S. terrestris
We have previously observed that B. subtilis (referred to as Bs ALBA01) isolated from the rhizosphere of onion plants inhibits the growth of the soil fungal pathogen S. terrestris17. After 15 days of interaction with the soil fungal pathogen S. terrestris in coculture assays, B. subtilis acquired strong antifungal activity. Such activity was mediated by secreted factors, since bacterial cell-free supernatants displayed the fungal growth inhibition17 (Fig. 1a, e). These Bs ALBA01 variants were termed post-ST, to distinguish them from pre-ST variants grown without any contact with the fungus and whose cell-free supernatants displayed no antifungal activity17 (Fig. 1a, e). Comparisons between pre-ST and post-ST variants revealed clear phenotypic differences, with post-ST variants showing greater degrees of roughness and wrinkled colony phenotype, thicker and more structured pellicles associated with robust biofilm formation and, lack of swarming motility (Fig. 1b–e).
Fig. 1: Coculture of B. subtilis with S. terrestris induces stable phenotypic changes. a Growth inhibition of S. terrestris at day 4 after inoculation. Strong growth inhibition was observed in dishes containing only cell-free supernatants of post-ST. b Phenotypic changes of Bs ALBA01 resulting from interaction with fungus on LB agar culture. Colonies of post-ST were rougher and more wrinkled than those of pre-ST. Agar plates (panels 1 and 4) and individual colonies (panels 2 and 5) of pre- and post-ST, respectively. Electron microscopy revealed increased numbers of elongated cells and reduced numbers of sporulated cells in post-ST (panel 6) relative to pre-ST (panel 3). Scale bar: 2 µm. Arrows: sporulated cells. c Post-ST variants in liquid medium developed thick, wrinkled pellicles characteristic of strong biofilm formation. Quantification of biofilm formation on borosilicate surfaces by crystal violet staining revealed greater biofilm formation by post-ST than by pre-ST. Data shown are mean values of measurements of optical density at 595 nm of crystal violet suspensions from three independent replicate experiments. ***Significant difference between values (p < 0.0001, Tukey’s multiple comparison test). d Swarming motility assessed on 0.7% agar LB plates. Pre-ST displayed full motility (i.e., covered the entire plate), whereas post-ST showed no swarming motility (i.e., did not extend beyond the inoculation area). e All B. subtilis variants obtained following coculture with S. terrestris acquired similar stable phenotypes. The panel shows phenotypes of other two post-ST, post-ST variants 2 and 3, and variants whose genomes were sequenced. Similarly to observations of post-ST, post-ST2 and 3 showed rough, wrinkled colony appearance, and robust pellicle formation relative to pre-ST. Swarming motility was absent in post-ST2 and 3. Dishes containing only cell-free supernatants of post-ST2 and 3 showed strong growth inhibition of S. terrestris (7 days after inoculation, at 30 °C). Full size image
Importantly, all post-ST phenotypic traits were maintained when variants were reinoculated onto the fresh growth medium (passaging), indicating that B. subtilis experienced genetically stable phenotypic variation upon coculture with S. terrestris.
Post-ST variants differ in metabolome and reduced surfactin
We next characterized the metabolites profile changes suffered by B. subtilis due to conversion to post-ST, which might be responsible for the antagonistic effects observed on S. terrestris. Thus, by using nontargeted metabolomics profiling based on 1H-NMR spectroscopy, we analyzed cell-free supernatants from pre- and post-ST. Pre-ST samples clustered together and were clearly separated from post-ST samples (Supplementary Fig. 1a, c, d). S-line plots in orthogonal partial least squares discriminative analysis (OPLS-DA) models showed clearly distinct signals, representing discriminant metabolites between the pre- and post-ST sample groups (Supplementary Fig. 1b).
Furthermore, we also compared chemotypes of whole cells in search of significant metabolomics differences between the two variants. Nontargeted high-performance liquid chromatography tandem mass spectrometry (HPLC–MS/MS)-based metabolomics analysis revealed separation of global metabolome between the two variants in principal coordinates analysis (PCoA; Fig. 2a). Surprisingly, the two well-known Bacillus lipopeptides, surfactin and plipastatin, were severely reduced in the post-ST variant (Supplementary Fig. 2a, b). This finding was intriguing because surfactin and plipastatin are two of the most important compounds employed by B. subtillis to thrive in microbial antagonistic interactions19,20. In fact, random forest analysis showed that they were the most important variables determining differences between pre- and post-ST (Supplementary Fig. 2c). Post-ST showed notable reduction in intensity of surfactin ions and of peaks with masses indicative of plipastatin, identified through spectrum library matching and feature-based molecular networking3,21,22. Importantly, post-ST clustered together with a B. subtilis NCBI 3610 surfactin-defective mutant23, confirming the role of surfactin in separation between the groups (Fig. 2a). Molecular networks obtained by analysis of chemical profiles of the two variants indicate reduction in post-ST of not only surfactin and plipastatin derivatives, but also related unknown derivatives (Supplementary Fig. 3). We further complemented this analysis with HPLC–MS/MS-based metabolomics analysis of cell-free supernatants of the two variants. Once again, PCoA plots evidenced a discrimination according to the origin of cell-free supernatants, displaying separate clustering of pre-ST vs. post-ST samples (Fig. 2b). Consistently with results of whole-cell analysis, supernatants of post-ST did not show surfactin (Fig. 2c) or plipastatin (Fig. 2d) ions levels comparable to those of pre-ST. Accordingly, post-ST supernatants showed no hemolytic activity (Supplementary Fig. 2d).
Fig. 2: LC–MS/MS-based metabolomics analysis of whole cells and cell-free supernatants showing chemical signatures that distinguish pre- from post-ST variants. a PCoA plots show strong separation between whole-cell sample analysis of pre- and post-ST. b PCoA plots show clear distinctions according to the origin of cell-free supernatants, with separate clustering of pre- vs. post-ST samples. Post-ST variants clustered together and separately from the pre-ST ancestral variant. The group of post-ST samples in PCoA plots included samples of three independently obtained post-ST variants (post-ST1, post-ST2, and post-ST3) after coculture with S. terrestris in vitro (see “Methods” for details). Metabolomics experiments were carried out in triplicates for each post-ST variant. c Intensities of surfactin and d plipastatin ions were strongly reduced in cell-free supernatants of post-ST. Full size image
Since the absence of surfactin in post-ST was unexpected, we decided to evaluate the role of surfactin deficiency on B. subtilis in this interaction process. To examine the possibility that key genes involved in surfactin production are related to acquisition of antifungal activity by post-ST variants, we generated a surfactin-defective Bs ALBA01 mutant (hereafter Bs srfAA) by disrupting the surfactin synthase encoding gene srfAA, essential component of the machinery for non-ribosomal synthesis of surfactin. Anti-S. terrestris activity of this mutant was screened upon 15-day of coculture by obtaining and testing samples before and after coculture. Interestingly, the absence of surfactin resulted in the acquisition of equivalent mycelial growth inhibitory activity to that observed in post-ST. In fact, pre- and post-ST supernatants of surfactin-defective Bs srfAA showed no notable differences in antifungal activity (Supplementary Fig. 4). Thus, inability to produce/release surfactin resulted in enhanced anti-S. terrestris activity, as well as complete suppression of ST-driven transformation to post-ST phenotype. On the other hand, unlike the post-ST variants and in agreement with previous reports6, the Bs srfAA strain lost the capacity to form biofilms.
2-Ketones mediate growth inhibition of S. terrestris
What two vegetables burn fat overnight?
Spinach and other leafy green vegetables like kale, lettuce, etc. are great for burning belly fat and are very nutritious as well. There have been...
Given that metabolomics profiles from HPLC–MS/MS showed clear differences, it was intriguing that the main difference in the metabolome of post-ST variants was the suppression of lipopeptides surfactin and plipistatin, whereas no candidate antifungal metabolites showed significantly higher abundances. However, our HPLC–MS/MS analysis was biased toward a limited area of the chemical space and was unable to detect small primary metabolites and volatile compounds. In order to expand our search to small molecules, we analyzed the volatile compounds profile of the variants and their interaction with S. terrestris. Thus, we adapted the growth of cocultures to glass vials suitable for headspace GC–MS analysis using solid-phase microextraction (SPME). From PLS-DA, we found that 2-heptanone was the most discriminating compound produced in greater amount post-mutation only by B. subtilis, and not by the fungus. Interestingly, by performing a GC data molecular network analysis24 within the Global Natural Products Social (GNPS) platform22, we could discriminate a cluster composed of a family of ketones, remarkably, all 2-ketones (Fig. 3a). Several of these ketones were also among the top discriminant features between pre-ST and post-ST variants and S. terrestris (Fig. 3b). Interestingly, individually cultured pre-ST variants showed high levels of 2-ketones (Fig. 3b). However, coculture with S. terrestris was able to induce a steep drop in 2-ketones levels in pre-ST, but remarkably post-ST variant retained the ability to produce them. Thus, we wondered whether 2-heptanone and a representative compound of the ketone cluster, 2-octanone, were sufficient to exert antifungal activity on S. terrestris. To the best of our knowledge, there is no specific genic pathway described in B. subtilis responsible for the synthesis of 2-ketones and thus, no 2-heptanone- nor 2-octanone-deficient mutants of B. subtilis can be achieved. Therefore, we decided to test the antifungal effect of either 2-ketones on S. terrestris. The activity of both compounds was tested in bioassays by culturing the fungus on potato dextrose agar (PDA) dishes containing a filter paper disc, where different concentrations of each purified compound were loaded. The growth of S. terrestris was gradually reduced as the concentration of 2-heptanone and 2-octanone increased, and the fungal inhibition was maximum at concentrations of 0.02 and 0.006 M, respectively (Fig. 3c). These results indicate that 2-heptanone and 2-octanone were involved in the antagonistic process against S. terrestris.
Fig. 3: Volatile compounds change due to interaction between B. subtilis and S. terrestris: 2-ketones induce fungal growth inhibition. a GC molecular network analysis revealed that Bs ALBA01 produces and releases a family of 2-ketones compounds, which arise as the most important volatile metabolites produced only by the bacterium. b Interaction with the fungus (ST) generates a steep drop of 2-heptanone (2H) and 2-octanone (2 O) levels in pre-ST, but not in post-ST. Both 2H and 2O are the most discriminating features between pre- and post-ST during coculture with ST. Boxplots show compound abundances, normalized by quantile normalization and auto-scaled (zero-centered and divided by SD). c Growth inhibition of S. terrestris at day 7 after inoculation of 0.02 M 2H and 0.006 M 2O on a filter paper disc placed on PDA dishes. Graphs showing the gradual reduction in mycelium growth with increasing concentrations of 2-ketones. Data shown are mean values of mycelial growth from three independent replicate experiments; red arrow indicates lethal concentration of 2H and 2O at day 7 after inoculation. Full size image
Surfactin suppresses the antifungal activity of 2-ketones
Considering our HPLC–MS/MS results, the antifungal behavior of the Bs srfAA mutant strain described above and the lethal effect of both volatiles 2-heptanone and 2-octanone on S. terrestris, we next explored whether the presence of surfactin was involved in the lack of antifungal activity characteristically observed in cell-free supernatants of pre-ST variants. If this were the case, then surfactin depletion would be a functional trait of post-ST variants necessary to achieve antifungal capacity. To evaluate this, we added exogenous purified surfactin to cell-free supernatant of post-ST and tested the anti-Setophoma activity of this combination. Exogenous surfactin partially suppressed the antifungal activity of post-ST supernatants (Fig. 4a). Furthermore, in order to distinguish if this suppressor effect was the outcome of surfactin directly interfering with the activity of antifungal compounds, we tested if surfactin was able to decrease the anti-Setophoma activities of 2-heptanone and 2-octanone. The exogenous addition of surfactin protected S. terrestris from the antifungal activity of both, 2-heptanone and 2-octanone (Fig. 4b and Supplementary Fig. 5).
Fig. 4: Surfactin interferes with the anti-S. terrestris activity of 2-ketones. a Suppressor effect of exogenous surfactin (1 mg/ml) on the antifungal activity of cell-free supernatants of post-ST observed at day 5 after inoculation. b Suppressor effect of a filter paper strip imbibed with surfactin (1 mg/ml) on the antifungal activity of 2H (0.02 M) and of 2O (0.006 M), 7 days after inoculation. Data shown are mean values of mycelial growth from three independent replicate experiments; red arrow indicates lethal concentration of ketones at day 7. Statistically significant differences at p < 0.0001, p < 0.001, and p < 0.05 are identified by ***, **, and *, respectively (one-way ANOVA followed by Tukey’s multiple comparison test). Full size image
Taking into consideration that post-ST variants can maintain effective levels of 2-ketones when interacting with the fungus, we suggest that surfactin interferes with 2-ketone activity, probably by direct chemical reaction and/or by physical sequestration through micelle formation. Alternatively, some authors hypothesize that surfactin may have a stabilizing effect on the lipid bilayers in the fungal membrane generating some kind of resistance to anitifungal compouds13. Yet, the underlying mechanisms of surfactin interference with 2-ketones or other compounds remain unexplored and need to be further investigated.
In line with our results, previous reports proposed that surfactins could interfere with the activity of other lipopeptides with documented antifungal activity13,25. By studying the efficacy of Bacillus amyloliquefaciens strains at inhibiting Rhizomucor variabilis, a fungal pathogen of maize plants, Zihalirwa Kulimushi et al.13, showed that fengycin and/or iturins, but not surfactin provided the main antifungal potential against R. variabilis. However, they revealed that the coproduction of surfactin together with fengycin decreased the global antifungal activity. Interestingly, the authors speculate that fengycins may be involved in permeabilization of spore/conidia and, therefore, inhibiting germination and/or causing hyphal cell perturbation. Opposing this, surfactin may have a stabilizing effect on certain lipid bilayers, thus limiting pore formation in fungal membranes. They also propose that under certain conditions and concentrations, surfactins and fengycins may co-aggregate, and form inactive complexes. Moreover, Kim et al.25 reported that a surfactin-deficient mutant of B. subtilis subsp. krictiensis ATCC 55079 showed increased levels of lipopeptide iturin A and higher antifungal activity against the fungus Fusarium oxysporum, compared to its parental wild-type strain. The authors suggested that higher antifungal activity was achieved not only by the enhanced production of iturin A, but also by the absence of surfactin, which could be either blocking the iturin A production by sequestrating substrates and/or by changing the activity of positive/negative regulators of cyclic lipopeptides.
Post-ST variants show loss-of-function mutations in QS genes
What food has the most fiber to help you poop?
Foods containing high levels of soluble fiber include dried beans, oats, oat bran, rice bran, barley, citrus fruits, apples, strawberries, peas,...
Finally, in order to elucidate the genetic bases of the phenotypic changes underwent by B. subtillis upon interaction with S. terrestris, we sequenced whole genomes of three post-ST variants (post-ST1, post-ST2, and post-ST3; Fig. 1e) derived independently from the Bs ALBA01 strain, which was considered as pre-ST ancestor and used as reference. Reads from these three variants were aligned against Bs ALBA01 genome26 (NCBI Bioproject PRJNA316980) to assess genetic changes accumulated in post-ST during interaction with the fungus. We found one to three different mutations in coding sequence regions for each of the variants (Supplementary Table 1). Interestingly, each variant had at least one loss-of-function mutation in genes of the ComQXPA QS system. The post-ST1 and post-ST2 variants had a frameshift mutation in coding sequence of the comA gene, while post-ST3 variant harbored a 100 bp deletion in the kinase-encoding gene comP. Mutations in comA and comP were further confirmed by PCR amplification and sequencing.
Then, we investigated the extent to which com genes in Bs ALBA01 were convergently mutated upon interaction with S. terrestris by performing whole-genome sequencing of a pool of 15 post-ST variants obtained independently, following interaction with the fungus in coculture (Pool-seq; see “Methods”). To identify allelic variations in the pool, we mapped sequence reads to the B. subtilis ALBA01 reference genome, and then performed variant calling (see Supplementary Methods). Of note, six different loss-of-function mutations were identified in the comQXPA operon within the pool of post-ST variants, revealing marked convergence of mutation within these genes (Supplementary Table 1). These findings demonstrate the occurrence of mutation-based phenotypic changes in post-ST variants and support the notion that interaction of B. subtilis with S. terrestris induces metabolomics changes promoting bacterial antifungal activity, which involves the ComQXPA QS system. In addition, they suggest that single mutations in one of the ComQXPA genic components provided pleiotropic effects for the adaptation of B. subtilis to antagonistic interactions.
The ComQXPA system has been reported to be required for transcription of numerous genes involved in competence development, antibiotic production, exopolysaccharide production, degradative enzyme production and transport, and fatty acid metabolism27,28. In fact, over 10% of the B. subtilis genome is controlled by the ComQXPA QS system28. Mutations in ComA could possibly provoke alterations on the general metabolism promoting chemical conditions that favor the biosynthesis of 2-heptanone and 2-octanone, as those produced through the degradation of branched-chain amino acids, leucine, valine, and isoleucine, which could otherwise be used for surfactin or lipopetides synthesis. Alternatively, it has also been described that ComA controls the expression of FapR, a transcriptional regulator involved in fatty acid synthesis in B. subtilis28. In this sense, FapR negatively regulates the expression of at least ten genes (the fap regulon)29. Thus, mutations in comA/comP could determine an increased fatty acid biosynthesis by downregulating FapR. The release of the fap regulon in association to the chemical condition derived from the global change produced by the mutation may be favoring branched-chain fatty acid biosynthesis pathways and/or decarboxilation of intermediates ketoacids that promote the synthesis of 2-ketones. Accordingly, a recent report showed an enhancement of de novo fatty acid synthesis in Bacillus nematocida B16 during 2-heptanone production30.
Notably, ComA activates transcription of the srfA operon that governs surfactin production31,32. The ComQXPA system has also been reported to play a role in transcription of regulator DegQ, which controls the production of degradative enzymes and is required for plipastatin production33,34,35. Thus, mutations in ComQXPA are in agreement with the loss of surfactin and plipastatin production observed in post-ST.
Until recently, it was widely accepted that production of surfactin was essential for biofilm formation. Thus, the enhanced biofilm formation displayed by post-ST variants resulted at first contradictory. However, it has been recently described that B. subtilis strains carrying mutations in genes of the ComQXPA QS system actually form pellicles with more matrix components than the wild-type strain36. In fact, ComQXPA mutants are able to develop earlier and thicker biofilms even though with decreased surfactin production. In agreement with these antecedents, we hypothesize that post-ST variants carrying loss-of-function mutations in ComQXPA system, promote the biofilm formation as an adaptive strategy to overcome interaction with the fungus. In addition, another study which described the diversity of ComQXPA among isolates from the tomato rhizoplane, also revealed a remarkable diversity in the surfactin production, which was not always positively correlated with the biofilm formation37. The authors observed that srfA gene inactivation in different natural isolates of B. subtilis showed different effects on biofilm biomass production, either displaying lower, equal, or higher biofilm formation than the wild-type strain. These observations indicate that the regulatory role of surfactin in biofilm formation may vary among different natural isolates of B. subtilis.
Among many other functions, it has been shown that inactivation of QS can lead to a delay in sporulation entry36, which is in accordance with our observation of a reduced sporulation level observed in post-ST variants (Fig.1b). Interestingly, it has been discussed that ComQXPA system may act as a switch that contributes to the stochastic initiation of sporulation which, consequently, is able to achieve a bet-hedging behavior to limit the investment in population growth, and favor commitment to late growth adaptive processes36.
Keto Means
Photo: Kevin Gamba
Promotion
Promotion
Promotion
Promotion